The biomedical measurements such as electrocorticography and electromyography have been widely performed for medical monitoring, diagnosis, and treatment of the diseases. In particular, subdural electrodes directly placed on the surface of a brain have been developed for a treatment of epilepsy and surgery. Conventional subdural electrodes consisting of metal electrodes such as Pt and the silicone-based substrate has various problems; the silicone substrate is not able to completely adhere to the brain surface due to its hardness and hydrophobicity, and artifacts will be generated in MRI imaging due to metal electrodes. We have developed a total organic hydrogel-based subdural electrode by embedding carbon fabric (CF) modified with poly(3,4-ethylenedioxythiophene) (PEDOT) into a poly(vinyl alcohol) (PVA) hydrogel substrate. Their softness similar to the living tissue and conformability to the curved surface of brains enabled easy handling, and the total organic materials will contribute to clearer MRI images without image artifacts. Moreover, molecules can pass through the hydrogel substrate of the electrode. Currently, we are working on the development of a hydrogel salt bridge electrode that is more transparent than conventional and our reported electrodes and reduces adverse effect on the body by avoiding the electrode from the biotissue surface and embedding salt bridges in the hydrogel substrate (Figure 1a). Although silicone microchannel is needed for a separation between hydrogels for the salt bridge and substrate, the bonding strength of silicone and hydrogel is poor, which causes a slip of the electrodes from measuring points. Various techniques for chemical bonding of silicone and hydrogel have been reported [1] [2], however, some of the substances used in these techniques are toxic and not biocompatible. In this study, we employed a physical bonding method with protrusions which are placed on the silicone surface to reduce the sliding of silicone and hydrogel. A frustum structure was selected as the shape of the protrusion, as for avoiding an exfoliation of the PVA hydrogel substrate from the silicone microchannel (Figure 1a).In order to evaluate the bonding strength, we prepared micro-structured polydimethylsiloxane (PDMS) sheets with 3D printed molds. The various micro-structured PDMS sheets were designed as rectangular (2 mm in the width) and frustum protrusions (4 mm, 2 mm, 1 mm and 0.5 mm in width) (Figure 1b). PDMS sheets with protrusions were embedded in PVA gel (15wt%), which were cross-linked by repeating freezing and thawing cycles (10 min at -30 °C and 10 min at room temperature for three times). A digital force gauge and an electric stand were used to perform tensile tests. Tensile loads were applied to both ends of the specimens, and the tension was recorded until the PDMS sheet and PVA gel were completely separated. In order to apply to the electrocorticography, we prepared the salt bridge electrode combined with the protrusions. The electrode was fabricated by injecting PVA precursor (10wt%) into the PDMS channel with protrusions and embedding that in PVA gel (15wt%). Each PVA was cross-linked by repeating freezing and thawing cycles at the same condition the above experiment. Then the electrode was placed on the brain of a pig along with a conventional electrode, and the brain wave was measured by electrocorticography (ECoG).Among the prepared PDMS sheets, the bonding strength of specimens with frustrum protrusions (10.8 kPa for 2 mm) showed higher than that of specimens with rectangular protrusions (9.3 kPa) and without protrusions (8.0 kPa) (Figure 1c). The bonding strength increased with a decrease of the size (0.5 mm, 2 mm and 4 mm) of protrusions (Figure 1d). Only 1 mm of protrusions did not follow the manner, which was under discussing. These results show that the frustum protrusions contribute to the physical bonding. Finally, we measured the ECoG signal by using the prepared hydrogel electrode. The salt bridge electrode was highly flexible enough to closely contact with the surface of the brain. Moreover, the ECoG signal measured by the hydrogel electrode was followed to conventional electrodes (Figure 1e). In conclusion, we successfully fabricated a flexible and biocompatible subdural electrode.